Energy and bandwidth optimized prefetching for mobile video streaming in CCN

Background
Mobile video streaming is one of the most energy consuming smartphone applications. In order to cope with bandwidth fluctuations, video stream players prefetch content. Prefetching large chunks of content is also energy efficient because of so called tail energy which is a fixed amount of energy wasted for each chunk downloaded irrespective of its size. However, more often than not the user abandons viewing before the end especially when viewing user-generated content (e.g. YouTube). Therefore, too aggressive pre-fetching leads to energy wasted for unnecessary download of data. It is also harmful because of unnecessary use of bandwidth. Therefore, optimizing the prefetching requires some estimate of when the user will abandon viewing, i.e. what is the internal popularity of the content being streamed[1].

Content Centric Networking (CCN)[2] is a pretty recent revolutionary concept for computer networking. In a nutshell, the idea is to completely change the paradigm from networking between named end hosts to networking named content. In practice, it means that in addition to the nodes that publish content, routers in the network cache it, and end hosts running applications request pieces of content directly from the network without caring which node in the network will serve it. While changing the whole Internet to function as one big CCN is an unrealistic goal in short term, there is growing interest in smaller scale deployments, for instance, by mobile network operators.

Because of the in-built caching mechanisms, CCN knows about the content popularity. A mobile client can potentially leverage that knowledge in order to optimize the prefetching of video streaming content.

Objectives and description of work
The goal of this thesis is to design and implement a CCN proxy/gateway functionality which enables the mobile client to request information about the video internal popularity in order to make optimized decisions about prefetching. The first part of the work focuses on designing the necessary CCN related non-trivial mechanisms so that video popularity information can be efficiently tracked. The second part of the thesis work is to implement and prototype of the designed solution and test it. The testing could be done with a small scale CCNx[3] test bed that could be deployed behind the Otaniemi campus LTE network (NetLeap).
 

Skills Needed

• Basic understanding of optimization techniques
• C/C++ programming skills

Contact :  Jussi Kangasharju,  Jussi.Kangasharju@helsinki.fi

References

1. Mohammad Ashraful Hoque, Matti Siekkinen, Jukka K. Nurminen. Using Crowd-Sourced Viewing Statistics to Save Energy in Wireless Video Streaming. In The 19th ACM Annual International Conference on Mobile Computing and Networking (MobiCom). October 2013.
2. V. Jacobson, D. K. Smetters, J. D. Thornton, M. F. Plass, N. H. Briggs, R. L. Braynard (PARC) Networking Named Content, CoNEXT 2009, Rome, December, 2009.
3. CCNx: http://www.ccnx.org/ 
4. Reinhard Grandl, Kai Su, Cedric Westphal, On the Interaction of Adaptive Video Streaming with Content-Centric Networking, 2013.
5. Christos Tsilopoulos, George Xylomenos and George C. Polyzos, Are Information-Centric Networks Video-Ready?, The Packet Video Workshop 2013.